So, the other answers have established that rowCount() shouldn't be used to count the rows of a SELECT statement. The documentation even says, that :
PDOStatement::rowCount() returns the number of rows affected by the last DELETE, INSERT, or UPDATE statement executed by the corresponding PDOStatement object.
So it's okay for other queries, but not great for SELECT. Most answers suggest that you should make two queries, one to count rows, and one to get the subset of records you need. However, you could query the row count and your subset of the data in one request. This is a bit of an exercise in code golf, but could actually prove more efficient than two requests if the request time is a bit costly and these requests are made frequently.
If you're in PostgreSQL you can provide clean JSON output, like so:
WITH mytable as (VALUES(1,2,3),(4,5,6),(7,8,9),(10,11,12))
SELECT
jsonb_build_object(
'rowcount', (SELECT count(1) FROM mytable)
,'data', (
SELECT jsonb_agg(data.*)
FROM (
SELECT *
FROM mytable
WHERE column1 > 1 -- pagination offset
ORDER BY column1
LIMIT 2 -- page size
) as data
)
) jsondata
Output:
{"data": [
{
"column1": 4,
"column2": 5,
"column3": 6
},
{
"column1": 7,
"column2": 8,
"column3": 9
}
],
"rowcount": 4
}
If you're not in postgres, those functions won't be available, but you could do this:
WITH mytable as (VALUES(1,2,3),(4,5,6),(7,8,9),(10,11,12))
SELECT
(SELECT count(1) FROM mytable) as rowcount
,data.*
FROM (
SELECT *
FROM mytable as mytable(column1, column2, column3)
WHERE column1 > 1 -- pagination offset
ORDER BY column1
LIMIT 2 -- page size
) as data
but it will return the rowcount on every row, which might be a bit wasteful:
| rowcount |
column1 |
column2 |
column3 |
| 4 |
4 |
5 |
6 |
| 4 |
7 |
8 |
9 |